Perioperative Pain Management in Newborns

The practice of medicine has become progressively more sophisticated. Facilitating the healing process while simultaneously minimizing or even eliminating pain once thought necessary to achieve healing is now an attainable goal. Adults were the first to benefit from these advances. Only within the past few decades has the sophisticated medical establishment realized that pediatric patients, including neonates, also feel pain and require medical intervention to alleviate unnecessary suffering. ^{[1, 2, 3]}

Beliefs of caregivers substantially affect medical intervention to alleviate unnecessary suffering. Before the late 1980s and early 1990s, a common belief was that neonates experienced no pain or less pain than adults, children, or infants who underwent similar surgical procedures. A healthcare provider who believes that neonates feel little pain tends to observe few clinical signs of pain in neonates. Furthermore, such practitioners tend to believe that narcotic administration is associated with increased risk in neonates. Although these practitioners may administer narcotic analgesics, they frequently use nonaggressive intervention and subtherapeutic regimens in preverbal patients compared with adults.

A concern about respiratory depression is the most notable limitation to the administration of narcotics postoperatively, especially in nonintubated neonates or neonates undergoing minor surgical procedures. Although this concern may have some pharmacologic basis, it should not prevent the appropriate administration of narcotic analgesics to neonates who have undergone clinically significant surgery. Furthermore, other therapeutic regimens formerly reserved for adults, adolescents, and older children may also be used safely to manage postoperative pain in neonates.

This article considers the application of medical advances in pain management to the care of neonatal surgical patients.

Pain is minimized during the preoperative visit by avoiding unnecessary laboratory studies, especially those that require phlebotomy. Furthermore, inappropriately long periods of taking nothing by mouth (NPO) can be eliminated to minimize patient discomfort. Postoperative pain management should be discussed during the preoperative visit with the family. Issues that may eventually affect decisions about postoperative pain management should be addressed, including the patient's coexisting disease states, the surgical site, the patient's postoperative disposition, and the family consent for pain-management techniques that under consideration.

Neonates who are unstable, septic, or likely to remain intubated after surgery are frequently given narcotics during the procedure. These narcotics are continued as needed afterward. Narcotic administration is cautiously implemented in neonates who are postoperatively transferred to non-ICU settings.

Neonates undergoing outpatient surgery or surgery associated with minor postoperative pain are frequently given acetaminophen with regional or local anesthetic infiltration after surgery. Neonates who undergo lower-extremity, abdominal, or thoracic surgery are excellent candidates for regional anesthesia, whether they are undergoing inpatient or outpatient surgery. As in adults, coexisting pulmonary disease in neonates may be an excellent reason to consider regional anesthesia for postoperative pain management.

Finally, anxiety among family members concerning a pain-control strategy should be thoroughly addressed and considered in decision-making about postoperative pain management.

After extensive work in the 1980s and 1990s, the fact that neonates experience pain and mount a stress response was established and appreciated. Metabolic and hormonal indicators of the degree of stress a surgical patient has can be monitored during and after surgery. These indicators are perioperatively elevated in neonates. Even premature neonates undergoing surgery can mount a clinically significant stress response, as measured by hormonal and metabolic indicators. Stress indicators include plasma adrenaline, noradrenaline, glucagon, insulin, and cortisol levels as well as blood glucose, lactate, pyruvate, and alanine concentrations.

The mounting of a surgical stress response results in catabolic responses, including glycogenolysis, gluconeogenesis, and lipolysis during the perioperative period. These catabolic responses, when unmodulated by medical intervention, may have a detrimental effect on the clinical course of a neonatal surgical patient.

Adverse circulatory and respiratory events are also most likely to occur during the postoperative course of neonates who have had inadequate interventions to minimize stress responses. Tachycardia, systemic hypertension, pulmonary hypertension, respiratory embarrassment, and intraventricular hemorrhage may be associated with inadequate pain control in neonates. Furthermore, inadequate treatment of pain may have implications that extend beyond the neonatal period, including hypersensitivity to noxious stimuli later in life.

Management of the surgical stress response in neonates can largely be accomplished with the same pharmacologic interventions as those used for anesthesia in other surgical patients. ^[4] Volatile anesthetic agents are the most common means of providing intraoperative anesthesia and analgesia probably because they meet, at least to some degree, each of the criteria required for complete anesthesia, including some hypnosis, amnesia, analgesia, and muscle relaxation.

Anand et al demonstrated that blood levels of hormonal and metabolic indicators of the stress response were lower in neonates who received volatile anesthetics during surgery than in those who did not. ^[5] Furthermore, clinical stability of neonates during and after surgery improved with the adequate administration of volatile anesthetic agents during surgery.

The relative potency of each volatile anesthetic agent is measured in terms of the minimum alveolar concentration (MAC) of the inhaled agent at which 50% of patients do not have skeletal muscle movement in response to surgical incision or another noxious stimulus. The patient's age appears to influence the MAC of a given volatile anesthetic. Although the MAC is higher in infants than in any other age group, it may be 15-25% lower in neonates than in infants and even lower in premature neonates.

Volatile anesthetics include potent myocardial depressants and vasodilators. As a consequence, systolic blood pressure and mean arterial blood pressure may decrease when these agents are administered. In some neonates, other analgesic agents may be used to decrease the requirement for volatile anesthetics during surgery to avoid some of the hemodynamic changes that may occur with the administration of these agents. In fact, surgical anesthesia can be accomplished without the use of any volatile anesthetic agents.

Narcotics are not complete anesthetic agents because they do not provide muscle relaxation or amnesia, which are essential components of complete anesthetics. However, they are potent analgesic medications. In neonates, blood pressure may decrease less with intraoperative use of narcotics than with the administration of volatile anesthetics. Furthermore, narcotics are useful adjuncts to anesthetics based on volatile agents because they can reduce the requirement for volatile agent, reducing any hemodynamic lability. Narcotic-based anesthetics are commonly used during cardiovascular procedures in neonates.

The pharmacokinetics of narcotics administered to neonates differ from the pharmacokinetics of narcotics given to infants, children, and adolescents. Compared with older patients, neonates have a lower rate of clearance, increased volume of distribution, prolonged elimination half-life, and heightened plasma concentration after a bolus dose of narcotics. As a consequence, intraoperative use of narcotics may affect the postoperative disposition of neonates. Narcotics commonly used for intraoperative analgesia in neonates include morphine, fentanyl, sufentanil, and remifentanil.

Ketamine, a phencyclidine derivative, produces amnesia and intense analgesia. This drug affects opioid receptors and N -methyl-D-aspartate (NMDA) receptors, as well as voltage-sensitive calcium ion channels as it induces its analgesic effects. Ketamine actually stimulates the cardiovascular system; therefore, it is frequently associated with increased systolic and mean arterial blood pressures and heart rate when administered intraoperatively.

Ketamine may not be associated with increased blood pressure in preterm neonates. Instead, it is associated with decreased mean arterial and systolic blood pressures, although these decreases are smaller than those of the other analgesic medications commonly used during surgery. Beneficial effects of ketamine include bronchodilation and less depression of ventilation than that observed with other agents. Adverse effects in neonates include increased production of salivary and tracheobronchial secretions, cerebral vasodilation, and apnea in patients with increased intracranial pressure.

A software program developed in Brazil appears to have the potential to provide clinicians with real-time sensitive and specific neonatal pain monitoring in the NICU. ^[3] The application uses real-time analysis of neonatal pain-related facial movements while the infants undergo painful procedures.

One factor that has contributed to inadequate pain management in neonates is the pervasive belief that neonates do not feel pain. This misconception is perpetuated, at least in part, by the conspicuous absence of adequate tools to assess pain levels in this patient population. To a large extent, pain assessment in patients older than neonates depends on their ability to report pain level in some form to the caregiver. When patients cannot express pain verbally, pain assessment depends more on evaluations by the caregiver than by the patient.

Even when pain is evident, quantifying the level is not easy. An effective pain-assessment tool must enable healthcare providers to objectively quantify the patient's level of pain so that they can accurately measure the effectiveness of interventions designed to alleviate unnecessary suffering. Although no perfect tool exists for assessing pain in neonates, infants, and preverbal children, several useful tools are available.

The Children's Hospital of Eastern Ontario Pain Scale (CHEOPS) was one of the first observational pain scales. ^[6] This tool includes the categories of (1) cry, (2) facial expression, (3) verbal response, (4) torso position, (5) leg activity, and (6) arm movement in relationship to the surgical wound. In general, each category is scored 0, 1, 2, or 3, with high scores indicating high pain levels; the scale varies with each category evaluated. The CHEOPS was originally used to evaluate postoperative pain in children aged 1-7 years. Evaluators determined that it was both valid and reliable for assessing pain in this patient group. However, a pain-assessment tool that is appropriate for preverbal children may not be appropriate for neonates. Still, CHEOPS is a tool against which the validity of other pediatric pain-assessment tools can be measured.

The Objective Pain Scale (OPS), which Hannallah et al developed, is both valid and reliable in pain assessment. ^[7] The OPS is used to assess (1) blood pressure, (2) crying, (3) movement, (4) agitation, (5) posture, and (6) verbalization. Each parameter is scored 0, 1, or 2, with high scores indicating increased distress. This instrument is important because it includes a cardiovascular parameter in the assessment of postoperative pain. Many advocate use of cardiovascular parameters as the most objective measures of the pain response in preverbal children. However, their use is limited because other causes of distress may dramatically change these parameters. Although cardiovascular parameters are insufficient as the sole means of assessing pain in this patient population, they may be helpful. However, the OPS, like the CHEOPS, has been used predominantly for infants and children and not neonates.

The COMFORT scale has been favorably received as a tool to assess postoperative pain in the neonatal population. ^{[8, 9]} This tool was originally developed to assess distress in ventilated patients in the pediatric ICU. ^[10] However, the COMFORT scale was reliable and valid in assessing pain in postoperative patients in one large study in which pain was evaluated in 158 neonates along with older infants and children. ^[11]

The COMFORT scale comprises 6 behavioral items: (1) alertness, (2) calmness, (3) muscle tone, (4) movement, (5) facial tension, and (6) respiratory response, and (7) two physiologic items (heart rate and mean arterial blood pressure). Each item is scored 1, 2, 3, 4, or 5, with a high score indicating increased distress. The additional variables assessed and the increased number of scores possible for each variable may make this tool more useful than others in identifying subtle changes in patient discomfort. On the other hand, the relative complexity of this scale may be a disadvantage in terms of its clinical use.

The COMFORTneo scale, one of several modified versions of the COMFORT scale, was developed specifically for use with neonates. ^[12] Van Dijk at al evaluated the scale using nearly 3600 triple ratings for 286 neonates and found good internal consistency, validity, and interrater reliability. ^[13] Another modification, COMFORT Behavior scale, has also received favorable reviews. ^{[9, 14, 15]}

A fourth scale is the crying, requires oxygen, increased vital signs, expression, and sleeplessness (CRIES) scale. ^[16] This tool may also be useful to assess the pain of neonates after surgery. This scale is designed to analyze 5 variables: (1) crying, (2) requirement of increased oxygen administration, (3) increased vital signs, (4) expression, and (5) sleeplessness. Each variable is scored 0, 1, or 2. This instrument has demonstrated validity, reliability, user friendliness, and acceptance as a tool to assess postoperative pain among neonatal intensive-care nurses.

The FLACC observational pain tool has demonstrated validity and reliability as an instrument to measure the pain of nonverbal or preverbal surgical patients. The parameters observed when using this tool include (1) face, (2) legs, (3) activity, (4) cry, and (5) consolability. Each parameter is scored 0, 1, or 2, with a higher score indicating increased distress. A revised version of the FLACC tool was recently shown to demonstrate improved validity and reliability for pain assessment in cognitively impaired children. ^[17]

The neonatal pain, agitation, and sedation scale (N-PASS) has become one of the 5 most commonly used measures of pain in neonates. ^[18] This scale uses 5 parameters scored 0, 1, or 2 in each category. The categories include (1) crying/irritability, (2) behavior/state, (3) facial expression, (4) extremities/tone, and (5) vital signs (heart rate, respiratory rate, blood pressure, and oxygen saturation). However, this scale does not distinguish pain from agitation.

The PIPP scoring system measures 5 parameters including heart rate, oxygen saturation, and 3 facial actions to assess pain in premature infants. Behavioral state and gestational age are used to modify the ultimate PIPP score. This tool has been determined to have validity and reliability among those trained in its use for assessing procedural and postoperative pain in premature infants.

The Neonatal Infant Pain Score (NIPS) and the Neonatal Facial Coding System (NFCS) have been determined valid instruments for procedural pain assessment in neonates. They have been used for postoperative pain assessment in this patient group as well. Variables measured by the NIPS tool include crying, facial expression, breathing patterns, arousal, and arm and leg movement. The NFCS assesses facial actions.

Each of the pain-assessment instruments discussed has strengths and limitations. For optimal use of any pain assessment tool, the neonatal staff should select a tool, become familiar with its use, and systematically integrate its use into the institution's policies. The use of this tool in combination with the direct input of physicians, nurses, and parents gives the best opportunity for accurate assessment of the neonate’s pain state. This approach preserves the validity and reliability of the assessment tool in measuring pain in neonates while simultaneously allowing for appropriate intervention to be undertaken, which minimizes unnecessary postoperative pain in neonatal patients.

After appropriate pain-assessment practices are established, the most formidable hindrance to alleviating postoperative pain in neonates is unfamiliarity with the safety and practicality of the options. Implementing an effective pain-management strategy in the neonatal surgical patient is a complex process. The strategy for pain management should begin during the preoperative assessment and continue with the intraoperative anesthetic management, as formerly discussed.

Furthermore, the clinician responsible for pain management must be aware that such care affects other components of postoperative care. When treating pain in neonates, one must consider the pharmacodynamic and pharmacokinetic issues unique to the neonatal period, the severity of the surgical insult, the patient's coexisting diseases, the surgical site, the postoperative management plans of the surgeons and neonatal or pediatric specialists, and the neonate's postoperative disposition. ^[19]

Opioid administration remains the most common means of achieving pain control in surgical patients. Neonates who are expected to have moderate-to-severe postoperative pain are no exception. Opioids may be safely administered to neonates when a well-constructed pain-management plan is implemented. Although certainly not prohibitive, the risk of apnea cannot be ignored. As during the intraoperative period, more than one opioid may be considered to manage postoperative pain in the surgical neonatal patient.

Fentanyl and morphine are the drugs most commonly selected when opioids are administered after surgery.

Fentanyl

Fentanyl is most appropriately administered by means of intravenous (IV) infusion to neonates who are preoperatively ventilated and who are expected to remain ventilated postoperatively. Episodes of apnea are more severe when this drug is administered as an IV bolus of 1-2 mcg/kg than when it is given as a continuous IV infusion of 1-2 mcg/kg/h. However, respiratory depression may be less problematic in infants than in neonates. Too-rapid bolus administration has been associated with acute rigidity of the chest wall that severely compromises ventilation of the patient.

As with many other medications administered to neonates, the pharmacokinetics of fentanyl widely vary. Of particular interest, the half-life of the drug may be prolonged in neonates who have increased abdominal pressure after surgery because of impaired hepatic blood flow. Fentanyl is less likely to cause hypotension than morphine; this feature may be a consideration in the postoperative neonate.

Morphine

Morphine remains the opioid analgesic most commonly used to control moderate-to-severe postoperative pain in neonates. ^[20] Surgical procedures after which morphine is given may include craniotomy, thoracotomy, sternotomy, and laparotomy. Incremental IV boluses of 20 mcg/kg, not to exceed 100 mcg/kg, are typically administered for acute pain management in the postanesthesia recovery unit. When a continuous IV infusion is used for postoperative pain management in neonates, the initial rate varies depending on the patient's age.

Initial IV infusion rates of 10 mcg/kg/h are acceptable for neonates younger than 1 week. Neonates older than 1 week tolerate 15 mcg/kg/h, whereas older infants may tolerate 20-40 mcg/kg/h. Supplemental IV boluses of as much as 50 mcg/kg may be administered for episodes of breakthrough pain in neonates who are receiving morphine by means of continuous infusion.

Respiratory responses to morphine do not differ in term neonates compared with infants and children when identical plasma levels of the drug are achieved and maintained. Increased caution may be advisable when morphine infusions are started in preterm neonates, who may require infusion rates significantly lower than those given to term neonates to achieve the same plasma levels of morphine. At steady-state plasma concentration of more than 20 ng/mL, the likelihood of respiratory depression increases in all neonates.

Practitioners who administer morphine infusions for postoperative pain control should be aware of the pharmacokinetic disadvantages that place neonates at risk for respiratory depression because of increased plasma concentrations. Neonates have immature hepatic enzyme systems, which may result in a doubling of the elimination half-life of morphine in neonates (ie, 10-20 h) compared with half-lives in infants (ie, 5-10 h). Low plasma protein levels in neonates may result raise levels of free drug and slower plasma clearance of morphine. Morphine clearance in neonates may be as little 50% of that of infants and 25% of that of adults. Furthermore, the rate of glucuronidation, the primary metabolic pathway of morphine, is slower in neonates than in infants and adults.

Coexisting surgical and medical conditions may affect the pharmacokinetics of morphine in neonates. As mentioned before, abdominal surgery increased abdominal pressure postoperatively may impair drug metabolism or drug elimination, increasing the half-life of this drug. In addition, the pharmacokinetics of medications may be different in neonates with cardiac comorbidity than in those without congenital heart disease. Whether this change is secondary to impaired cardiac performance or abnormal circulatory dynamics on hepatic or renal function is unclear. Use of preservative-free morphine should also be considered because neonates are more susceptible than other patients to respiratory depression some preservatives cause.

Morphine can also stimulate histamine release, resulting in systemic vasodilation. This characteristic may be either advantageous or disadvantageous to the postoperative neonate, and it should be considered when this drug is used.

Meperidine

Meperidine is seldom used for postoperative management of pain in neonates, although it can be administered in doses of 1 mg/kg. However, its pharmacokinetic profile dramatically varies in individual neonates. Furthermore, meperidine has epileptogenic metabolites, which tend to accumulate in patients with impaired renal function, as is inherent in the neonatal period. In the authors' opinion, the disadvantages of meperidine outweigh its advantages in the neonatal patient.

Nonnarcotic analgesics

Nonnarcotic modalities are important in pain management in neonates who have undergone surgery. Acetaminophen is useful as a sole analgesic for mild discomfort or as an adjuvant medication for moderate-to-severe pain when narcotic or regional analgesia is used. ^{[21, 22]} Either oral or rectal administration may be chosen. Rectal administration is associated with low plasma levels and a prolonged elimination half-life. Although 15 mg/kg may be administered orally, this dose is associated with subtherapeutic plasma levels when administered rectally to neonates. An initial dose of 20-35 mg/kg is recommended for the initial preincision dose or for the immediate postoperative dose administered per rectum.

In general, the risk-benefit profile for nonsteroidal anti-inflammatory drugs (NSAIDs) does not favor their use in neonates. Potential complications include renal toxicity, necrotizing enterocolitis, platelet dysfunction, and hemorrhage (particularly intracranial hemorrhage). In particular, the US Food and Drug Administration (FDA) has not approved ketorolac for use in neonates, and reports of its use in this patient population are absent from the literature. Moreover, ketorolac may be no more effective than high-dose rectally administered acetaminophen in some older patients.

Sucrose 25% administered orally and repeatedly can blunt the pain response in neonates undergoing common but painful procedures such as circumcision, heel lancing, and venipuncture. Doses of 2-3 mL may be administered intraoperatively into the oral cavity by using a syringe or by using a pacifier dipped in the solution and then inserted into the patient's mouth. A commercially prepared solution is available, or it can be mixed at the bedside by using 1 packet of table sugar dissolved in 5 mL of water. Other sugars, such as dextrose and glucose, have not been shown to have the same analgesic effect.

Nonpharmacologic interventions

A discussion of nonnarcotic pain modalities is incomplete without a discussion of nonpharmacologic interventions. Although some form of these modalities may be used in older patients, they are considered central to the pain management of neonates. ^[23] These modalities include bundling, holding, and rocking the neonate; provision of a pacifier to alleviate distress; and minimization of environmental stimuli, such as extraneous noise and unnecessary light.

Regional pain-control techniques are increasingly used to manage postoperative pain in neonates. Regional techniques may include single-dose administration of local anesthetics into the caudal space, plexus blockade of the upper or lower extremity, extrapleural catheter placement, or neuraxial catheter placement for continuous pain control after surgery. The most common regional techniques in neonates include single-dose caudal administration and placement of epidural catheters for prolonged pain management. ^{[24, 25]}

Caudal anesthesia is a highly effective and simple technique associated with a high success rate and a low complication rate. Caudal anesthesia is neuraxial anesthesia and is therefore associated with some of the risks inherent to neuraxial access. However, because the neuraxial space is accessed at its most caudad entry point, the risk of neural injury or even inadvertent dural puncture is reduced. Sterile technique is required and may be accomplished by wearing sterile gloves or by palpating the caudal space anatomy through an alcohol swab (ie, the no-touch technique) before a single dose of medication is instilled into the caudal space.

Use a short, beveled needle to minimize the likelihood of inadvertent intravascular or intramedullary injection of the local anesthetic. A caudal anesthetic can be successfully administered in 96% of pediatric patients. After the sacrococcygeal ligament is penetrated with the regional anesthetic needle, lower the angle of the needle, advance it no more than 3-5 mm, aspirate the syringe to ensure the absence of cerebrospinal fluid (CSF) or heme, and administer the local anesthetic.

Bupivacaine remains a commonly administered local anesthetic for single-dose caudal blocks. ^{[26, 27]} However, newer local anesthetics are being used with increasing frequency. Ropivacaine is now an extremely commonplace local anesthetic for caudal blocks and may offer distinct advantages compared with bupivacaine because ropivacaine has a relatively benign cardiovascular toxicity profile. Levobupivacaine may offer the same advantage. The incidence of residual motor blockade is lower with both of these agents than with bupivacaine, without diminishment of analgesic efficacy. Effective concentrations are 0.125-0.25% for bupivacaine, levobupivacaine, and ropivacaine. Volumes of 0.75-1 mL/kg and/or doses of up to 3 mg/kg are typically administered. Supplemental analgesics may not be required for as long as 12 hours after surgery when the caudal block is effective.

Placement of a caudal, lumbar, or even thoracic catheter for continuous postoperative pain management has been proven safe and effective in neonates. Ultrasonography further enhances both the success rate and the safety of neuraxial anesthesia in pediatric patients. An epidural catheter may be successfully placed using the caudal approach and advanced cephalad to the lumbar or thoracic level. By using superficial anatomic landmarks as a guide, the level of the catheter may be accurately predicted. This catheter may then be used for a postoperative infusion of narcotics or local anesthetics.

Epidural catheters have been successfully used after many major neonatal surgical procedures that require laparotomy or thoracotomy, including hepatic resection and procedures to treat abdominal-wall defects (gastroschisis and omphalocele), tracheoesophageal fistula, congenital diaphragmatic hernia, and coarctation of the aorta. After successful catheterization and administration of aninitialbolus through the epidural catheter, pain management may be maintained with a continuous infusion of analgesic medications.

Epidural infusions are an acceptable alternative to the intermittent top-up technique. Epidural infusions are both safe and effective in term and preterm neonates. Postoperative epidural infusions of bupivacaine substantially reduce sedation and respiratory depression. Moreover, patients have improved oxygenation without the administration of supplemental oxygen with analgesia, complication rates, and hemodynamic profiles similar to that of patients who receive morphine infusions.

In 1992, Berde reported recommendations to facilitate the safe use of epidural analgesia in pediatric patients after analyzing results in more than 20,000 pediatric regional anesthetic procedures at 15 institutions. ^[28] One of Berde’s recommendations was that bolus dosing of epidural bupivacaine should not exceed 2-2.5 mg/kg. Infusion rates of 0.2-0.25 mg/kg/h were recommended for neonates. The report cautioned that children are probably not more resistant to local anesthetic toxicity than adults, as was previously thought. Neonates, in particular, may be at risk for local anesthetic toxicity because of diminished plasma levels of alpha1-acid glycoprotein, which could increase the free fraction and decrease the clearance of bupivacaine. Premonitory symptoms or signs of local anesthetic toxicity may be absent in neonates. Reduce infusion rates for patients at risk for seizures.

When the level of the epidural catheter is too low to provide adequate analgesia at the incision site in a neonatal patient, increasing the rate of the epidural infusion cannot safely overcome this problem. In one study, plasma levels of bupivacaine continued to increase and reached the upper limits of the safe range before the end of a 48-hour infusion period. The rate of increase in plasma levels of bupivacaine was most dramatic in neonates who had increased abdominal pressure after surgery. Furthermore, as with all drugs administered during the neonatal period, interindividual variability in plasma bupivacaine levels were considerable in neonates receiving epidural infusions. Although plasma clearance is lower in neonates than in adults who receive epidural infusions, this difference is especially dramatic in preterm neonates.